基于MSC Nastran的离散变量优化算法的实现

实际工程中存在大量的离散变量优化问题，基于MSC Nastran优化框架实现新的离散变量算法，有利于新算法本身的推广应用和解决大规模的实际复杂工程问题.通过修改MSC Nastran输入文件的方法实现离散变量的优化算法——GSFP算法.GSFP是基于广义形函数的离散变量优化算法，它将离散变量优化问题转化成连续变量优化问题，通过惩罚等措施使得最优设计结果最终收敛到离散解，该方法能够解决大规模的实际离散变量优化问题.最后以桁架截面选型优化为应用背景，给出GSFP算法实现的基本原理和方法.

Implementation of discrete variable optimization algorithm based on MSC Nastran

There are large number of discrete variable optimization problems in practical engineering. A new discrete variables algorithm can be achieved through the MSC Nastran optimization framework, which is conducive to the promotion and application of the new algorithm and to solve the large scale complex engineering problems. The discrete variable optimization algorithm GSFP algorithm by modifying the MSC Nastran input file is implemented. The GSFP is a new discrete variable optimization algorithm based on generalized shape functions, which can translate the discrete variable optimization problem into a continuous variable optimization problem and adopt punishment and other measures. So the optimal design results are eventually converged to the discrete solution. This method can solve large scale discrete variable optimization problems. The truss sectional selection optimization to introduce the fundamental and implementation methods of achieving the GSFP algorithm based on MSC Nastran is chosen.